Oil Price Shocks and Inflation Targeting in China

Yunqing Wanga, b, Linsen Yinc, Xinyu Suid and Wenjie Pane, *

aSchool of Finance, Shanghai Lixin University of Accounting and Finance, 995

Shangchuan Road, Shanghai, China;

bInstitute of Macroeconomy and Strategy, PICC Asset Management Co., Ltd., 1198

Century Avenue, Shanghai, China;

cSchool of Financial Technology, Shanghai Lixin University of Accounting and

Finance, 995 Shangchuan Road, Shanghai, China;

dSchool of Economics and Management, Shanghai Bangde College, 299 Jinqiu Road,

Shanghai, China;

eSchool of Statistics and Information, Shanghai University of International Business

and Economics, 1900 Wenxiang Road, Shanghai, China.

* Correspondence: hanspwj@126.com

Oil Price Shocks and Inflation Targeting in China

We develop a small open economy DSGE-based New Keynesian model

incorporating the demand of oil, to focus on whether Chinese central bank targets

core inflation or headline inflation including oil price inflation, as well as

investigating the macroeconomic effect of oil price shocks. Based on both

counterfactual simulations and welfare evaluations, our results indicate that

targeting both core inflation and non-core (i.e. oil price) inflation are inferior to

the one purely pegged to core inflation, suggesting the central bank should target

core inflation instead of headline inflation. Our paper contributes to a growing

literature on monetary policy in China and other emerging market economies.

Keywords: inflation targeting; core inflation; headline inflation; Chinese

monetary policy

Subject classification codes: E31; E43; E47

1. Introduction

As noted in Mishkin (2010), there were two global IT waves in the last 10 or so years of

the 20th Century: the information technology (IT) wave and the inflation targeting (IT)

wave in the context of monetary policy. Currently, the countries using relatively mature

inflation targeting policies, have made great achievements with reliable evidence. Not

only has the long-term inflation been significantly reduced, but the fluctuations in

inflation rates have also decreased (Bernanke 2001).

As the largest emerging market economy, in the process of reform and opening

over the past three decades, China’s inflation has undergone a transformation, moving

from high inflation to a soft landing, followed by deflation and then mild inflation.

During the deflationary period, from 1997 to 2002, a proactive monetary policy was

largely ineffective because financial invocation in China and the acceleration of the

opening up process led to an erratic money multiplier, which in turn decreased the

effectiveness of the money supply as an intermediate target. Subsequently, China’s

economy entered another phase of rapid growth, and although the PBoC continued to

increase its benchmark lending rate and statutory deposit reserve ratio, it was unable to

effectively curb the acute fluctuations in property prices or their continued overall rise.

Following the financial crisis in 2008, China’s contractionary monetary policy was

moderately eased, finally becoming stable. The broad money supply, or M2, grew at an

average of 18% per year, which was much higher than GDP growth rate, but the

injection of large amounts of money did not alleviate the slump in the real economy.

Although monetary credit is circulating internally, the proportion of credit circulating in

vain is growing, which greatly reduces the effectiveness of a monetary policy that has

money supply as its intermediate target. In such complicated context, exploring the

choices of inflation targeting sheds light on the design and implementation of optimal

monetary policy for Chinese economy.

According to the Law of the People’s Republic of China on the PBoC, the objective

of the monetary policy is to maintain the stability of the value of the currency and

thereby promote economic growth. In that case, optimal objective of monetary policy is

concerned with what inflation indicator to target, i.e., whether the focus should be on

headline inflation or core inflation. Headline inflation measures changes in the cost of

living and includes a weighted average based on an expenditure ratio of the prices of all

goods and services consumed by residents. The commonly used Consumer Price Index

(CPI) is an example of a headline inflation measure. According to Eckstein (1981) and

Bryan and Cecchetti (1994), headline inflation, which is represented by the CPI, can be

separated into two major components. One is the underlying rate determined by

aggregate demand and aggregate supply, which is known as core inflation. The other is

a temporary rate caused by particularly volatile items like food and energy, which is

known as non-core inflation. Because core inflation attempt to smooth volatile changes

in the CPI to distinguish the inflation signal from the transitory noise, it is viewed as the

long-term underlying trend indicator of the CPI.

Mishkin (2007, 2008) suggests that monetary policy should focus on core inflation

instead of headline one including transitory, highly volatile part. Dhawan and Jeske

(2007) build an energy economy NK-DSGE model incorporating durable goods

consumption, and they conclude it is better to use core inflation than headline inflation

in the context of alternative Taylor rule. Based upon the Ramsey optimal policy rules,

Kormilitsina (2011) verifies the negative impact of monetary policy on the economy

will be worsened by rising energy prices, suggesting that monetary policy need to be

targeted on core inflation. In the case of China’s food price, Hou and Gong (2013)

demonstrate targeting core inflation can significantly reduce the welfare loss in response

to an increase in food price.

In practice, even among countries that clearly practice inflation targeting, the

inflation indicators used differ greatly. The majority of countries continue to use

headline CPI, which includes short-term variable components, as their target. The use of

headline CPI shows that on this issue, theoretical knowledge and practice are not

aligned, and in formulating policies, each country also takes into consideration its own

circumstances before making certain trade-offs. Food and energy form a major part of

residents’ expenditure and can affect overall prices through transmission. Hence, the

monetary policy pegging headline inflation will not only help stabilize price fluctuations

but also reduce consumer welfare losses. Cat~ao and Chang (2015) show that the high

volatility of world food prices cause headline CPI inflation targeting to dominate core

CPI inflation targeting in the context of a small open economy. Anand and Prasad

(2010) establish a two-sector two-good closed economy NK-DSGE model with

financial frictions to find that a welfare-maximizing central bank should adopt targeting

headline inflation rather than the core inflation for emerging markets, which can also

help keep the credibility of monetary policy and stabilize inflation expectations. But on

the other side, the prices of such items are vulnerable to non-market factors such as

climate change. Therefore, the frequent adjustment of monetary policy will have a

negative impact on the economy due to the long lags between monetary policy actions

and changes in economic activity (Mishkin 2007).

There are very few studies that have concerned about the question of how monetary

policy makers should assess the efficacy of inflation targeting in China under both a

headline inflation targeting regime, and a core inflation targeting regime. As the

Chinese economy grows and matures, the PBoC will need to restructure itself to deal

with these types of important institutional issues. We develop a sticky price DSGE

model here that provides an analytical framework for assessing the implications of

alternative monetary policies in response to oil shocks, as well as to evaluate optimal

monetary policy in order to answer which inflation measure is the more appropriate

focus of policy for China. The paper makes three main contributions.

Firstly, due to the unique nature of China’s economy, this paper relies on the

Bayesian method to solve parameters for the following reasons. First, unlike developed

economies such as US, comprehensive quarterly data on major macroeconomic

variables are only available for China after 1992. Classical models using econometric

methods such as OLS or MLE are often not robust for lacking of macroeconomic data,

while Bayesian techniques are useful for short-time series problem. Second, there exist

statistical inconsistencies and a certain amount of measurement error in China’s

macroeconomic data, while the Bayesian method can treat as random disturbances that

can be given some prior information structure, thereby overcoming to some extent these

shortcomings. Finally, in a transition economy such as that in China, there are often

structural breakpoints in its operation, and it is easy for jumps to appear in the process

of generating macroeconomic data. By keep updating probabilistic inference of the

model parameters, the Bayes theorem fully accounts for the transience of the economy

in various regimes, thereby addressing structural changes in economic parameters.

Secondly, we provide a more analytical framework that can be used to compare the

pros and cons of two types of inflation for the PBoC to target. Moreover, two exercises

are conducted to test robustness from a welfare point of view: one is to cumulate the

sum of the impulse response function under “first moment”, and the other is to evaluate

the utility-based welfare metric under “second moment”. Various sensitivity

experiments have been done to all above results. Our findings suggest that China’s

monetary policy should focus on core inflation instead of headline inflation.

Finally, as the second largest oil consumer in the world today, China is fairly reliant

on imports. Oil prices volatility in China may be highly correlated with international

markets, which may influence each other. In that case, simply assuming that oil prices

in the Chinese domestic market are entirely dependent on changes in the international

market, i.e., that the price of oil in China is completely exogenous, somewhat lacks

adequate theoretical and evidentiary support. Thus, in our model, oil price is determined

endogenously beyond the traditional works of exogenous oil/energy price (Kormilitsina

2011; Finn 2000; Wang and Zhu 2015), so the model enables us to study various

channels through which shocks that cause oil price hikes and China’s other

macroeconomic variables interact.

In recent related work, Anand et al. (2015) give a similar debate over headline

versus core inflation targeting by introducing financial frictions. They conclude that in

the presence of financial frictions, headline targeting is superior, and point out this is

more relevant to emerging market countries in which many households are often credit

constrained. But in fact, for consumers in China there is relatively abundant supply of

credit, and in China's urban-rural dual structure the “credit constraints” are untenable.1

Our model adds oil as factor input in Cobb-Douglas production function (with labor)

different from a liner technology only including labor in their model. Contrary to the

conventional method of taking the oil price variable as exogenous process, we explicitly

model it endogenously with DSGE model. More importantly, by introducing inflation in

oil price into the Taylor rule to indicate non-core inflation, we investigate more

intuitively the “pros and cons of headline and core inflation” in extended Taylor rule.

The paper is organized as follows. The next section outlines the structure of the

SOE model. Section 3 reports on estimation exercise and model parameters. Section 4

performs dynamic analysis of the model and moreover section 5 discusses which

inflation to target. Section 6 concludes the paper.

2. The Small Open Economy Model

In this section, we build a small open economy (SOE) DSGE-based New Keynesian

model based on the work of Galí and Monacelli (2005) (GM thereafter) and Unalmis et

al. (2009). Specifically, we assume that the world economy is composed of a domestic

SOE (like China) and a continuum of other small open economies (the rest of world, or

ROW), all distributed on [0, 1] uniformly. China and the ROW are completely identical

in terms of consumer preferences, firm technologies and market structure including the

Calvo price-setting mechanism and the assumption of complete financial markets. The

representative household consumes domestic and imported goods, supplies labor, earns

1 Credit constraint and even financial constraints seem to be an obvious, unquestionable proposition for a relatively less-developed countries and regions and intuitively people

think that this conclusion is remarkable. But in fact, for China, credit restraint itself is lacking of adequate theoretical and empirical evidence. On the one hand, according to the

survey and empirical research on Chinese farmers' borrowing behavior Zhong et al. (2010) found, the majority of households credit demand can be met. On the other hand, since

the late 1990s, the Chinese government has been expanding domestic demand and consumption, as one of the "Troika" for boosting economic growth, is highly expected during

the process of expanding domestic demand. In this context, consumer credit for improving consumer environment and expanding consumer is developing rapidly in support of

the government. Since the new century, the scale of Chinese consumer credit is developing at an average annual growth rate of 29 percent. By 2014, China's consumer credit

reached 15.4 trillion. China has become the largest country in Asia excluding Japan in the size of consumer credit, while consumer credit would weaken consumer precautionary

savings or myopic consumer motivation (Berg 2013), which is the premise of motivation for consumer credit constraints.

wages and shares dividends derived from the firms. Firms produce differentiated goods,

decide on labor and oil inputs, and set prices according to Calvo mechanism. The

government sets fiscal and monetary policy.

2.1. Households

A representative household is infinitely-lived and seeks to maximize

E0∑t=0

∞

β t [U (C t )−V ( N t ) ], (1)

U (C t )=log (Ct ), (2)

V ( N t )=N t

1+ψ

1+ψ, (3)

where N t denotes hours of labor and C t is a composite consumption index defined by

C t=1

(1−α )1−α αα CH , t1−α CF ,t

α,

where CH ,t and CF ,t are CES consumption indices of domestic and foreign goods, given

by :

CH ,t=(∫0

1

CH ,t (i)ε−1

ε di)ε

ε−1 ; CF ,t=(∫0

1

CF ,t (i)ε−1

ε di)ε

ε−1, (4)

where i∈¿ indicates the goods varieties, ε>1 is the elasticity of substitution among

goods produced within a country, α∈(0,1)represents the expenditure share of the

imported goods in households consumption basket, and is thus a natural index of

openness. ψ is the inverse of the elasticity of labor supply.

The maximization (1) is subject to a budget constraint of the form

Pt C t+Et {Q t ,t+1 Dt+1 }≤ Dt+W t N t+T t (5)

where Pt=PH , t1−α PF, t

α is the consumer price index (CPI) and the price indices for

domestically produced and imported goods are

PH ,t=(∫0

1

PH ,t (i)1−ε di)

11−ε ; PF,t=(∫

0

1

PF ,t (i)1−ε di)

11−ε ,

where Dt+1 is the nominal payoff in period t+1 of the portfolio held at the end of period

t including the shares in firms, Qt ,t+1 is the stochastic discount factor, W t is the nominal

wage and T t is lump-sum transfers or taxes.

Solving the household’s utility-maximization problem gives a labor supply

equation and an intertemporal Euler equation, respectively:

C t N tψ=

W t

Pt, (6)

β R t Et {(Ct+1

C t )−1

( Pt

Pt+1 )}=1, (7)

where Rt−1=Et {Qt ,t+1 } is the return on a riskless bond paying off one unit of domestic

currency in period t+1. Notice that (6) and (7) can be respectively written in log-

linearized form as:

w t−p t=ct +ψ nt (8)

c t=Et {ct+1 }−(rt−Et {π t+1 }−ρ ) (9)

where lowercase letters denote the logs of respective variables in uppercase letters,

ρ=−logβ, π t=p t−p t−1 is the CPI inflation between t and t+1.

2.2. The Terms of Trade, Inflation, Real Exchange Rate and UIP Condition

The (log) terms of trade, i.e. the price of foreign goods in terms of home goods, is given

by

st= pF ,t−pH ,t. (10)

Log-linearization of the CPI formula around a symmetric steady state satisfying the

purchasing power parity (PPP) condition PH ,t=PF ,t yields

pt=pH , t+α s t. (11)

It follows that domestic inflation, πH , t ≡ pH ,t−pH , t−1, and CPI-inflation are linked

according to

π t=π H ,t +α Δ s t. (12)

In addition, we define the real exchange rate Qt ≡Ξ t P t

¿

Pt, where Ξ t is the nominal

exchange rate (foreign currency in terms of home currency) and Pt¿ is the foreign price

index (where a star denotes foreign variables or parameters henceforth). Combining (10)

and (11), we derive the log-linearized formula for the CPI, domestic price level and real

exchange rate

pt=pH , t+α

1−αq t. (13)

Under complete international financial markets assumption, a Euler equation

analogous to (7) must also hold for consumers in the foreign country,

β ( Ct+1¿

Ct¿ )

−1

( Pt¿

Pt+1¿ )( Ξ t

Ξt+1 )=Qt , t+1. Together with (7), it follows (after iterating) C t=κ Ct¿Q t

around symmetric equilibrium where κ is a constant that depends upon initial

conditions, the log-linearized version of the risk sharing equation can be written as

c t=c t¿+q t.

Combine Euler equations of both countries to yield the uncovered interest parity

condition (UIP)

rt−rt¿=Et {Δe t+1 }, (14)

where e t ≡ log Ξ t−logΞ . Finally, we can derive the terms of trade in terms of the UIP

condition as

st=Et {∑k=0

∞

[ (r t+k¿ −π t+k +1

¿ )−(r t+ k−πH , t+ k+1 ) ]}. (15)

2.3. Firms

The home final good can be obtained by assembling intermediate goods varieties,

indexed by i∈¿: Y t=[∫01

Y t (i )ε−1

ε di ]ε

ε−1, where ε is the elasticity of substitution between

domestic varieties. Minimizing the cost of producing the aggregate implies that the

demand for each variety is given by Y t (i )=( PH ,t (i )PH , t

)−ε

Y t.

Intermediate good i∈¿ is produced by a monopolist according to the Cobb-Douglas

function:

Y t (i )=( A t N t (i ) )η (Otd (i ) )1−η, (16)

where η is the share of labor in the production, so that the share of oil is1−η. Here,

Otd (i ) is the amount of oil consumed in production by firm i, and the (log) labor

productivity a t=log ( A t ) follows a stationary AR(1) process, i.e. a t=ρa at−1+ϵ a , t, where

ρa is the parameters of persistence and ϵ a , t N ( 0 , σa ).

Under the assumption that firms take the price of each input as given, cost

minimization of the firm implies

(1−η ) (1−τ ) W t N t (i )=ηPto Ot

d (i ), (17)

MC t=(1−τ )

W t

PH , t

η Atη ( N t (i ) )η−1 (Ot

d ( i ) )1−η, (18)

where Pto denotes the nominal oil price which is determined endogenously in our model,

while making ~P to=

Pto

Pt as its real price. τ is an employment subsidy rate designed to

allow the flexible price economy to be efficient.

We assume that firms set prices in a staggered fashion in the spirit of Calvo

mechanism. In particular, in each period, a firm faces a constant probability, 1−ξ of

being able to reoptimize its nominal price. Thus, ξ is the probability that firm does not

change its price. As shown in Gali and Monacelli (2005), the firm’s optimal price

setting strategy implies the following marginal cost-based (log-linearized) Phillips

curve.

πH , t=β Et {π H ,t }+ (1−ξ ) (1−βξ )ξ

mct (19)

2.4. Goverment

We assume that the governments of both domestic and foreign countries are home-

biased. Similar to consumption index, the government spending index can be written as

Gt=(∫0

1

Gt (i )ε−1

ε dj )ε

ε−1 Moreover, expenditure minimization leads to the government

demand equation: Gt ( j )=( PH ,t ( i)

PH ,t)−ε

Gt. Moreover, the government follows a balanced

budget policy and finances its expenditures by lump-sum taxation, T t=PH ,t Gt . Also,

(log) government spending gt=log (Gt ) follows a stationary AR(1) process, i.e.

gt=ρg gt−1+ϵ g , t, where ρg is the parameters of persistence, the random error ϵ g ,t is i.i.d.,

and satisfies ϵ g ,t N (0 , σ g ).

Xie and Luo (2002) applied Chinese monetary policy to test Taylor rule, regarding

that Taylor rule can serve as a measure of China's monetary policy. Based on this, while

drawing researches of Clarida et al. (2001), Mohanty and Klau (2005), Mei and Gong

(2011), Zhao et al. (2016), the paper set China’s Taylor interest rate rule (log-linearized)

as follows:

rt=ρr rt−1+ (1−ρr ) [ϕπ π t+ϕx x t ]+υt (20)

where x t= y t− y t ( y t denotes output under flexible prices) is output gap. Assume ρr, ϕ π

and ϕ x are non-negative coefficients associated with interest rate rule. Note that interest

rate shock follows a stationary AR(1) process, i.e. υt= ρυ υt−1+ε υ, t where ρυ is the

parameters of persistence, the random error ϵ υ ,t is i.i.d., and satisfies ϵ υ ,t N (0 , συ ).

2.5. Equilibrium Dynamics

2.5.1. Domestic Demand Determination and IS Curve.

Lets CH ,t¿ (i ) denote the world demand for the domestic good i, i.e. export for domestic

country. Then market clearing in the small economy requires

Y t ( i )=CH , t ( i )+C H ,t¿ (i )+Gt ( i )

¿( PH ,t (i )PH ,t

)− ε

[( PH , t

Pt)−1

(1−α ) Ct +( PH ,t

Ξt P t¿ )

−1

α ¿Ct¿+G t ] (21)

Using Y t=(∫0

1

Y t ( i )ε−1

ε dj)ε

ε−1 and the law-of-one-price, we obtain

Y t=Gt+(PH ,t

Pt)−1

Ct (22)

Substituting it into the Euler Equation (9) of consumers, together with Equation

(13), one can write the domestic output curve (IS Curve)

y t=Et {y t+1 }−(1−G y ) (r t−E t {π t+1 })−G y Et {g t+1−gt }

−(1−G y ) α1−α

E t {q t+1−qt }(23)

where G y=GY denotes share of government spending in output.

2.5.2. World IS Curve and Aggregate Supply.

Market clearing in the world economy requires

Y t¿=C t

¿+Gt¿

(24)

Then the log-linearized version is y t¿=(1−G y) c t

¿+G y gt¿. Similarly, we can obtain IS

curve of the world output

y t¿=Et { y t+1

¿ }−(1−G y) (r t¿−Et {π t+1

¿ })−G y E t {gt+1¿ −g t

¿} (25)

Based on staggered price setting Calvo mechanism, the dynamics of inflation in the

world economy is

π t¿=β Et {π t+1

¿ }+ (1−ξ¿ ) (1−β ξ¿ )ξ¿ mct

¿ (26)

2.6. Oil market equilibrium and determination of endogenous oil price

In recent years, oil coming from abroad has accounted for an increasing proportion of

China’s aggregate amount. Chinese external dependence of petroleum and crude oil

both broke the point of 55% in 2011, surpassed the U.S. as the highest in the world.

Thus, it is possible that oil price volatility in China may be highly correlated with

international markets, implying that they react upon each other. In that case, it is not

appropriate for us to assume that China’s oil price is exogenous and entirely dependent

on changes in the international market. Therefore, departure from traditional literature,

we treat oil prices endogenously as follows.

Assume that at each point in time, the international oil supply Ots is exogenous and

its log-linearized variables follow AR(1) process such that o ts=ρo ot−1

s +εo ,t, of which ε o ,t

is i.i.d. and satisfies ε o ,t N (0 , σ o ) 2. As a small economy in the model, China’s oil

demands are relatively small compared to the ROW. To simplify, we assume that the

market-clearing condition of the world oil market is that oil demand equals oil supply,

i.e. o td∗¿=ot

s¿. Together with the world labor supply condition, the world production

function, cost minimization of the world firm and clearing conditions in the world

economy, as well as real world oil prices as ~p to∗¿=pt

o∗¿−pt¿¿ ¿, we can obtain the world oil

price expression (after log-linearization):

~p t

o∗¿=[ 11−G y

+1+ψ ¿

η¿ ] yt¿−

G y

1−G ygt

¿−(1+ψ ¿ ) at¿−[ 1+ψ ¿ ( 1−η¿)

η¿ ]ots¿. (27)

Using Ξ t P to∗¿=Pt

o¿ and the definition of real exchange rate, we obtain the real price

of oil endogenously:

~p to=~pt

o∗¿+ qt ¿(28)

In recent years, one focus of the research into international oil prices has been the

re-examination of structural reasons for fluctuations in oil prices (Hamilton 2008; Kilian

2009). A series of important studies in (Kilian 2008; Kilian 2009) indicate that the

causes are not limited to supply side, but also closely linked to macroeconomic factors

related to demand side. This proposition can be verified using the above two equations

(27) - (28). On the one hand, on the supply side, oil prices in China are inversely related

to the world oil supply; that is to say, as oil supply increases, oil prices will decrease,

which is in line with the classic supply and demand theory. On the other hand, from the

demand side, positive world government expenditure and labor technology shocks will

2 For instance, certain unforeseen political events (e.g. Iran-Iraq War, Gulf war) in OPEC countries cause dramatic changes in the world oil supply. This scenario supports for the

exogenous assumptions about the world oil supply. We assume that all the consumers equally share the profits of international oil supply in the world, same as Campolmi

(2008), Unalmis et al. (2009).

give rise to a decrease in domestic oil prices, implying that they share an inverse

relationship. Also, we found that an increase in world output and real exchange rates

pushes world real oil prices up.

3. Estimation of Parameters

Bayesian techniques are the principal strategy of solving parameters. However, if we

resort to Bayesian estimation to all parameters, some cannot be identified (Canova and

Sala 2009; Iskrev 2010). Therefore, we follow Smets and Wouters (2003, 2007) by

using Bayesian approach to estimate most other parameters after calibrating a few non-

identified parameters.3

Table 1. Parameter Estimates

Parameters Description Prior DistributionPosterior

mean

5%-95%

Posterior Interval

σ a China’s productivity shock INV_GAMMA[0.8, ∞] 1.1710 [0.95681.3711]

σ a∗¿¿ World’s productivity shock INV_GAMMA[0.8, ∞] 0.2898 [0.1848,0.3860]

σ g China’s gov. spending shock INV_GAMMA[4.0, ∞] 7.6256 [6.3650,8.9067]

σ g∗¿ ¿

World’s gov. spending

shockINV_GAMMA[4.0, ∞] 3.9959 [3.1646,4.7512]

σ υ China’s interest rate shock INV_GAMMA[0.5, ∞] 1.3154 [0.8822,1.7230]

σ υ∗¿¿ World’s interest rate shock INV_GAMMA[0.5, ∞] 0.2128 [0.1082,0.3100]

σ o

International oil supply

shockINV_GAMMA[5.0, ∞] 5.3464 [4.5218,6.1980]

α Degree of Openness BETA[0.40, 0.1] 0.2024 [0.1255,0.2791]

ψInverse elasticity of China’s

labor supply NORMAL[3.0, 0.75] 4.8252 [3.8553,5.8157]

ψ¿ Inverse elasticity of world’s NORMAL[3.0, 0.75] 4.3014 [3.3879,5.2538]

3 Because of length limitations, we omit the description of specific steps, information that is available upon request from the authors.

labor supply

ξ China’s calvo prices BETA[0.75, 0.05] 0.4339 [0.3872,0.4716]

ξ¿ World’s calvo prices BETA[0.75, 0.05] 0.6550 [0.5827,0.7272]

ρr

Coeff. on lagged Interest

rate of ChinaBETA[0.50, 0.1] 0.5486 [0.4318,0.6744]

ϕ π

Weight on inflation in

China’s Taylor rule NORMAL[1.75, 0.1] 1.6805 [1.5121,1.8426]

ϕ x

Weight on output gap in

China’s Taylor ruleNORMAL[0.70, 0.1] 0.6746 [0.5125,0.8382]

ρr¿

Coeff. on lagged Interest

rate of worldBETA[0.80, 0.1] 0.7170 [0.6175,0.8202]

ϕ π¿

Weight on inflation in

world’s Taylor ruleNORMAL[1.5, 0.1] 1.5067 [1.3442,1.6768]

ϕ x¿

Weight on output gap in

world’s Taylor ruleNORMAL[0.50, 0.1] 0.5342 [0.3710,0.7034]

ρa China’s productivity shock BETA[0.85, 0.075] 0.8170 [0.7573,0.8836]

ρa∗¿¿ World’s productivity shock BETA[0.85, 0.075] 0.8397 [0.7317,0.9506]

ρg

China’s gov. expenditure

shockBETA[0.85, 0.1] 0.9148 [0.8903,0.9395]

ρg∗¿¿

World’s gov. expenditure

shockBETA[0.85, 0.1] 0.3768 [0.2465,0.5031]

ρυ China’s interest rate shock BETA[0.85, 0.1] 0.1032 [0.0829,0.1242]

ρυ∗¿ ¿ World’s interest rate shock BETA[0.85, 0.1] 0.7683 [0.6314,0.9149]

ρo

International oil supply

shockBETA[0.85, 0.1] 0.7080 [0.5494,0.8722]

We find that economic openness parameter, α=0.2024, is estimated to be smaller

than the assumed in the prior distribution (0.4), implying that since China adopted the

reform and opening-up policy in 1978, it has rapidly integrated the world economy

globalization with rising openness. However, compared to advanced economies, it still

has a large room for increasing economy openness in future.

The estimate of ψ is 4.8252, which is greater than 4.3014 of the world, implying a

small estimation of the elasticity of labor supply to China. Chinese labor supply is also

relatively insensitive to wages, probably because as a country with the largest

population in the world, China has an abundant of labor supply.

The estimate of the price stickiness, ξ (0.4339), is smaller than the mean assumed

in the prior distribution, which means that prices are re-optimized frequently once every

two quarters. This is quite consistent with the market-oriented reforms of Chinese

government in recent years. Chinese government accelerates the market-oriented reform

on energy, transportation, electricity, communications and other large monopoly

enterprises, which is gradually decreasing the degree of stickiness in the entire society.

Finally, we move now to the parameters regarding monetary policy rule. The

posterior mean ρr of the interest rate smoothing coefficient is 0.5486, which is higher

than the prior mean. Interest rate policy of China has showed greater consistency since

the RMB exchange rate reform in 2005. Since then, the PBoC has more consideration to

the interest rate smoothing in formulating interest rate policies, to avoid the excessive

disturbance of capital market caused by abrupt changes in interest rates.

Meanwhile, we find a relatively small response to inflation and output gap (

ϕ π=1.6805 and ϕ x=0.6746) with respect to their prior distributions, indicating that

monetary policy in terms of interest rates rule operates insufficiency response. One

possible reason could be that effects of a quantity rule on Chinese economy seem to

have more significant than those of a price-based rule. This result is in line with the

findings of empirical studies on China’s monetary policy rules, reported by Xie and Luo

(2002) and Zhang (2009).

4. Dynamic Responses to Shocks

4.1. Impulse response analysis

As Figure 1 shows, note that although a reduction in international oil supply (a negative

shock) leads directly to a rise in oil prices in both China and the ROW, this type of

increase in oil prices have endogenous properties, which contrasts with a traditional

exogenous increase in oil prices. The oil markets clearing condition o td∗¿=ot

s¿, which was

used in the model mentioned earlier, indicates that exogenous shocks of international oil

supply firstly deliver to the ROW; therefore, let us start by analyzing the situation in the

ROW (or world). According to the production function, a reduction in the international

oil supply would cause a decrease in oil input, which in turn causes a decline in

production. At the same time, under the marginal cost function, a negative shock causes

the marginal costs of firms to rise, resulting in higher inflation. According to the Taylor

rule, central banks would then raise the nominal interest rate to combat inflation. As for

China, due to the exchange rate pass-through effect, rising real oil prices of the ROW

results in an increase in domestic real oil prices. On the one hand, this would cause a

decrease in firms’ demand for oil input, which in turn would lead to a decline in output.

On the other hand, this would cause an increase in marginal cost, resulting in higher

inflation and then causing the central bank to raise the nominal interest rate. Note that

according to the Taylor rule for both China and the ROW, the increase in the domestic

interest rate triggered by these shocks would be higher than that ROW because the

response coefficients of domestic inflation and output gap are larger than the ROW

ones. Coupled with uncovered interest rate parity, this scenario would appear an initial

appreciation in the nominal exchange rate followed by depreciation, which in turn

would lead directly to a similar trend in real exchange rate and terms of trade.

Figure 1. Impulse response of a negative standard deviation International oil

supply shock

4.2. Variance decomposition

As Table 2 shows, international oil supply shock mainly affects real oil prices, but has

little effect on the output and inflation, namely, no more than 10%, which is similar to

Zhao et al. (2016). Monetary policy shocks account for the most important sources of

inflation fluctuations, while labor productivity shocks have a big contribution to output

fluctuations in China, which is consistent with most of literature on RBC. Government

spending shocks account for 20.37% of the output variance, which is in line with

China’s recent dependence on increased government expenditure to boost the economy.4

Table 2. Variance Decomposition (Unit: %)

International

Oil supply

shock

China’s

Interest

Rate

Shock

World’s

Interest

Rate

Shock

China’s

Gov.

spending

Shock

World’s

Gov.

spending

Shock

China’s

productivity

shock

World’s

productivity

shock

Real Oil

Prices

69.83 0.13 9.40 7.31 9.51 3.69 0.13

Output 3.05 2.50 0.27 20.37 0.27 73.51 0.02

CPI-

Inflation

0.64 90.50 0.07 3.67 0.73 4.34 0.04

Inflation 0.79 81.76 0.42 5.98 0.33 10.69 0.03

Output

of the

ROW

8.60 0.00 24.87 18.46 28.74 0.00 19.34

Inflation

of the

ROW

0.75 0.00 91.99 2.75 3.73 0.00 0.77

4.3. Historical Decomposition

4 A typical example is the global financial crisis in 2008, which was triggered by the subprime mortgage crisis in the US. To stimulate economic growth, the Chinese government

rolled out fiscal expansion plans worth 4 trillion and implemented an expansionary monetary policy.

From Figures 2-1 to 2-3, histograms of various colors represent each type of shock, and

the black solid line represents the historical data. First, Figure 2-1 shows that real oil

prices in China have a strong positive correlation with international oil supply. For the

two sample periods of 2008Q2-2009Q2 and 2014Q3-2015Q3, the international oil

supply and world interest rate shocks play a major role in accounting for volatility in

real oil price. Second, Figure 2-2 shows that China’s output fluctuation is mostly driven

by government spending. During the global financial crisis from the second half of 2007

to the first half of 2009, the sharp decline in output was due mainly to the shocks of

interest rate policy and labor productivity. Finally, Figure 2-3 shows that most of the

variation in CPI inflation seems to be due to interest rate and government spending

shocks. In addition, Figures 2-2 and 2-3 indicate that international oil supply shock does

not contribute significantly to fluctuations in China’s output and inflation over time,

which is consistent with the above results from variance decomposition.

Figure 2-1. Historical variance decomposition of China’s real oil price

Figure 2-2. Historical variance decomposition of China’s output

Figure 2-3. Historical variance decomposition of China’s CPI inflation

5. What inflation indicator to target: core or headline inflation?

The simulation results show that rising oil prices trigger a loss in output and an increase

in inflation, which raise the risk of stagflation and therefore do have an adverse effect

on the economy. The “leans into the wind” feature of monetary policy may be required

to respond to oil price shocks, which raises the question that whether the PBoC should

take into account energy prices volatility represented by oil when executing monetary

policy. Since the financial accelerator theory on how changes in asset prices can amplify

macroeconomic fluctuations is proposed by Ben Bernanke, former chairman of the Fed,

and others, numerous academics have conducted successful research into the issue of

whether monetary policy must pay closer attention to fluctuations in asset prices. By

analyzing the relationship between the social welfare function and loss function of

monetary policy, it is discussed whether monetary authorities will take the specified

asset prices into account when executing monetary policy. Learning from their ideas, we

will take the oil prices as asset prices (Bernanke et al. 1997), introducing inflation in oil

prices as a variable in the Taylor rule in order to explore whether it is necessary for the

PBoC to consider oil price volatility when executing monetary policy. In summary, it is

to examine: what measure of inflation should the PBoC target, headline inflation

including oil prices or core inflation? The next part will answer these questions.

5.1. A numerical analysis of first moments

As mentioned above, the headline inflation can be decomposed into core inflation (trend

component) and oil prices inflation (temporary component). Similar to the setting of

Dhawan and Jeske (2007), intuitively, we use a simple linear equation, Π tHL=Π t + χe Π t

o,

to visualize, where Π tHL, Π t and Π t

o is headline inflation, core inflation and oil prices

inflation, respectively, and χe denotes the relative weight in headline inflation to oil

price inflation.

Due to the fact that oil price inflation (log-linearized) is defined as π to ≡ pt

o−p t−1o ,

we can modify the standard Taylor rule (equation (20)) as follows:

r t=ρr rt−1+ (1−ρr ) [ ϕπ π t+ϕπo π t

o+ϕx x t ]+υt (29)

where ϕ πo is the oil price inflation reaction parameter in the extended Taylor rule. When

ϕ πo=0, core inflation collapses to headline inflation, that is, Π t

HL=Π t , indicating that

monetary policy needs only target core inflation, rather than focusing on non-core

inflation determined by oil prices volatility. Leduc and Sill (2001) and Carlstrom and

Fuerst (2006) study the DSGE model of US oil economy in terms of this rule. When

ϕ πo >0, the monetary policy must take both core and non-core inflation into account,

which means monetary policy should target headline inflation. When ϕ πo <0, the

monetary authorities accommodate the oil price inflation, rather than the traditional

“headwind” regulation.

Hereafter, we adopt the counterfactual simulations to study the effects of different

weights on oil price inflation (ϕ πo)5 on output and core inflation in the event of an

increase in oil prices. Specifically, similar to the setting of Wang and Zhu (2015), five

sets of values ϕ πo ∈ [ 0,0.1,0 .2,0.5 ,−0.1 ] are selected to do the simulation, the first is the

benchmark value for the model; the following three are chosen in ascending principles,

a negative value is chosen as the last value to investigate the accommodating simulation

analog to oil price inflation (hereinafter referred to as the accommodating model), and

other deep parameters are assumed to be the same.

Figure 3 shows that as ϕ πo increases, a rising oil prices causes the loss in output in

the first period progressively. In particular, for the benchmark model of ϕ πo=0, if the

5 Note that there is no direct relation between the weight parameters χe and ϕ π

o . The former helps the reader to visualize, whereas the latter introduces oil price inflation into

the reaction coefficient of the Taylor rule equation under the monetary policy.

central bank follows the standard Taylor rule by selecting core inflation as the monetary

policy objective, then the loss in output in the first period would be less than that of the

other three model of ϕ πo >0, but more than that of accommodating model of ϕ π

o <0.

Though there is a slight reversal from the beginning of the second period, the output

values of all models in sample periods are negative. Furthermore, the cumulative

average output loss during the sample period (Ly) satisfies:

Ly+β L y+β2 L y+⋯+β19 Ly=∑t=1

20

β t−1 y t⇒ Ly=1−β20

1−β ∑t=1

20

β t−1 y t

and by contrast, it can be found the average output loss of 4.34% under the standard

Taylor rule is lower than three models’ value of the augmented Taylor rule ϕ πo >0, but

greater than that of the accommodating model.

Figure 3. Counterfactual Simulation of Rising Oil Prices (ϕ πo is variable)

Figure 4 shows that as ϕ πo increases, although a rising oil prices causes the decline

in inflation in the first period progressively, inflation starts to be increased from the

second period. Moreover, the overall price level changes (Δ pF) in the sample period is

calculated based on the core inflation rate, which is calculated as follows:

Δ pF=1+exp(∑t=1

20

π t)

Figure 4. Counterfactual Simulation of Rising Oil Prices (ϕ πo is variable)

By comparison with price volatility, the change rates in overall price level is 2.24 in

the last period with respect to the first period under the standard Taylor rule, which is

less than three models’ value of the augmented Taylor rule with ϕ πo >0, but higher than

that of the accommodating model.6

The objective of China’s monetary policy is to “maintain the stability of the value

of the currency and thereby promote economic growth”, indicating that stable inflation

and output promotion are core objectives in the implementation of its monetary policy.

In medium- to long-term, the standard Taylor rule, which results in a smaller decline in

output and a smaller increase in price levels after shock, is the preferred choice for the

PBoC. It also suggests the PBoC to target core inflation rather than headline inflation.6 We also make the robustness test, which can be requested from the author.

5.2. Numerical analysis of second moment

We follow Woodford (2003) and Galí (2008) by taking the second order approximation

of the utility losses of the domestic consumer resulting from shocks that hits economy.

Thus, as β → 1, the expected welfare losses of shocks can be written in terms of

variances of domestic inflation and output gap7:

V t=−(1−α )

2 [ ελ

var ( πH , t )+1+ψ

ηvar ( xt )].

Compared with GM model, the most prominent difference is the parameter of the

output gap volatility. Take the share of oil in the production equals to 0(η=1), which

indicates only labor input in the production function, thus the welfare loss function

collapses to that of GM. The existence of oil input, η(0<η<1), results in an increase in

the output gap volatility.

Making use of the above expression, we calculate the welfare loss in terms of

different interest rate rules for robustness test, and summarize the results in Table 3.

Regardless of what kind of deep parameters, the stronger the monetary policy response

to oil price fluctuations, the greater the corresponding welfare loss. This result indicates

that monetary policy pegged to headline inflation rate is not in favor of the improvement

of social welfare, on the contrary, the classic one that is only pegged to core inflation

rate could be considered as the best metrics for China's monetary policy, and can also

verify the conclusions of “the first moment” mentioned above.

Table 3. Contributions to welfare losses (international Oil supply shock with one

standard deviation)

Monetary Policy

7 Specific steps can be obtained from the author.

ϕ πo=0 ϕ π

o=0.1 ϕ πo=0.2 ϕ π

o=0.5

Benchmark model

Var(Domestic Inflation)) 0.0201 0.0258 0.0773 0.4343

Var(Output Gap)) 0.0010 0.0005 0.0047 0.0388

Total Welfare Loss 0.0671 0.0842 0.2600 1.4912

Strong nominal inertia ξ+20%

Var(Domestic Inflation)) 0.0158 0.0240 0.0648 0.3304

Var(Output Gap)) 0.0019 0.0009 0.0087 0.0720

Total Welfare Loss 0.0554 0.0794 0.2296 1.2383

Weak nominal inertia ξ−20 %

Var(Domestic Inflation)) 0.0239 0.0273 0.0895 0.5369

Var(Output Gap)) 0.0005 0.0002 0.0023 0.0190

Total Welfare Loss 0.0781 0.0883 0.2934 0.5369

Strong interest rules ( ϕπ+ϕx )+20%

Var(Domestic Inflation)) 0.0137 0.0157 0.0566 0.3546

Var(Output Gap)) 0.0007 0.0004 0.0042 0.0332

Total Welfare Loss 0.0458 0.0515 0.1922 1.2213

Weak interest rules ( ϕπ+ϕx )−20 %

Var(Domestic Inflation)) 0.0332 0.0514 0.1209 0.5534

Var(Output Gap)) 0.0015 0.0004 0.0050 0.0448

Total Welfare Loss 0.1104 0.1662 0.4009 1.8888

High degree of openness α +20 %

Var(Domestic Inflation)) 0.0201 0.0253 0.0759 0.4292

Var(Output Gap)) 0.0010 0.0004 0.0046 0.0381

Total Welfare Loss 0.0671 0.0823 0.2553 1.4731

Low degree of openness α−20 %

Var(Domestic Inflation)) 0.0200 0.0264 0.0788 0.4394

Var(Output Gap)) 0.0010 0.0005 0.0049 0.0395

Total Welfare Loss 0.0667 0.0861 0.2653 1.5093

6. Conclusions

Aggregate supply shocks, as represented by oil prices, present a challenging issue for

most countries. In the event of a severe shock that affects the economy, a monetary

policy response that targets a corresponding inflation indicator must be implemented.

Otherwise, inflation may exceed its target value, leading to costly losses in output and

employment. Therefore, both academic researchers and central banks give a rising

concern on the issue of whether to stabilize the headline or core inflation, also including

China. Being downturn brought about by economic restructuring, it is inappropriate for

China to select the money supply as the regulatory objectives of its monetary policy.

Learning from the international experience, inflation targeting is undoubtedly an

important direction of China’s deepening monetary policy institutional reform in future.

We use Chinese quarterly data to build New Keynesian DSGE model in the oil

economy. We utilize the state-of-the-art welfare evaluation method in modern monetary

economics to study the pros and cons of two types of monetary policies, namely

headline inflation targeting and core inflation targeting. Our purpose is to provide

theoretical support for future reforms to China’s monetary policy system.

1. Based on the calibrated model associated with seven structural shocks, the impulse

response function, variance decomposition, and historical variance decomposition

are calculated. Some conclusions are worth noting. First, International oil supply

shock plays a direct role in China’s oil price, whereas has a very limited impact on

China’s output and inflation. For the wide fluctuations of China’s real oil price

during the sample periods of 2008Q2-2009Q2 and 2014Q3-2015Q3, the

international oil supply shocks and the World’s interest rate shocks play leading

roles. Second, the interest rate policy shocks are the main driving force to push

China’s CPI Inflation, followed by government spending. Third, the government

spending shocks are the main strength which undertakes China’s output fluctuations.

However, during the global financial crisis starting from the second half of 2007 to

the first half of 2009, the dramatic decline of output growth rate was mainly due to

the combined shocks of China’s interest rate policy and labor productivity.

2. By integrated using counterfactual policies and welfare evaluations, our model

shows monetary policy that simultaneously targeting on core and non-core inflation

are inferior to the monetary policy that is purely pegged to core inflation, suggesting

the central bank should focus on core inflation instead of headline inflation in

setting monetary policy, thus providing a theoretical support for monetary policy

practice in the future.

In future work, it will also relax the implied assumption that the central bank has a

significant amount of credibility, to the point that temporary oil shocks do not spill over

to a permanent increase in prices and hence higher inflation over time8. The PBoC and

modern central banking is still in its infancy in China and will likely evolve

considerably in the coming years. To this end, as the institutions evolve and the

economy matures and becomes more open, such considerations of credibility in

maintaining its inflation target will become more important. As a result, it may be the

case that in order to establish its credibility, the PBoC may for some time need to target

headline inflation in order to anchor expectations permanently and establish credibility

before transitioning to a core targeting framework. Another extension would be to

include China’s prevailing policy regime such as capital control, exchange rate targets,

sticky wages, and many additional frictions in the model.

8 If the central bank had no credibility, headline and core inflation would essentially move in the same manner as we saw in the 1970s in the U.S., implying virtually no

difference in welfare between the two approaches.

Funding:

Please add: This research was funded by the Project of the National Social Science Fund

of China [grant numbers: 15CJY064].

Disclosure statement:

The authors declare no conflict of interest.

References

Anand, R.; Prasad, E.S. (2010). “Optimal price indices for targeting inflation under

incomplete markets.” NBER Working Paper, No.16290.

Anand, R.; Prasad, E.S.; Zhang, B.Y. (2015). “What measure of inflation should a

developing country central bank target?” Journal of Monetary Economics, 74, 102–

116.

Berg, E. (2013). “Are poor households credit-constrained or myopic? Evidence from a

South African panel.” Journal of Development Economics, 101(1), 195-205.

Bernanke, B.S. (2001). “Inflation targeting: lessons from the international experience.”

Princeton University Press: New Jersey, USA.

Bernanke, B.S.; Gertler, M.; Watson, M.; Sims, C.A.; Friedman, B.M. (1997).

“Systematic monetary policy and the effects of oil price shocks.” Brookings Pap Eco

Ac, 91-142.

Bryan, M.F.; Cecchetti, S.G. (1994). “Measuring Core Inflation.” In Monetary

Policy. The University of Chicago Press: Illinois, USA.

Campolmi, A. (2008). “Oil price shocks: demand vs supply in a two country model.”

MNB Working Paper, No.5.

Canova, F.; Sala, L. (2009). “Back to square one: identification issues in DSGE

models.” Journal of Monetary Economics, 56(4), 431-449.

Carlstrom, C.T.; Fuerst, T.S. (2006). “Oil price, monetary policy, and counterfactual

experiments.” Journal of Money, Credit, and Banking, 38(7), 1945–1958.

Catão, L.A.V.; Chang, R. (2015). “World food prices and monetary policy.” Journal of

Monetary Economics, 75, 69-88.

Clarida, R.; Galí, J.; Gertler, M. (2001). “Optimal Monetary Policy in Closed Versus

Open Economies: An Integrated Approach” NBER Working Paper, No. 8604.

Dhawan, R.; Jeske, K. (2007). “Taylor rules with headline inflation: a bad idea.”

Federal Reserve Bank of Atlanta, Working Paper, No. 14.

Eckstein, O. (1981). “Core inflation.” Prentice Hall: New York, USA.

Finn, M.G. (2000). “Perfect competition and the effects of energy price increases on

economic activity.” Journal of Money, Credit, and Banking, 32(3), 400-416.

Galí, J. (2008). “Monetary policy, inflation and the business cycle: an introduction to

the new Keynesian framework.”  Princeton University Press: New Jersey, USA.

Galí, J.; Monacelli, M. (2005). “Monetary policy and exchange rate volatility in a small

open economy.” Review of Economic Studies, 72(3), 707-734.

Hamilton, J.D. (2008). “Oil and the macro-economy.” In The New Palgrave Dictionary

of Economics; D.Stephen and B. Lawrence (eds); Palgrave MacMillan Ltd: New

York, USA; pp. 1–16.

Hou, C.Q.; Gong, L.T. (2013). “Food Price, Core Inflation and the Target of Monetary

Policy.” Economic Research(Chinese), 48(11), 27-42.

Iskrev, N. (2010). “Local identification in DSGE models.” Journal of Monetary

Economics, 57(2), 189-202.

Kilian, L. (2008). “The economic effects of energy price shocks.” Journal of Economic

Literature, 46(4), 871-909.

Kilian, L. (2009). “Not all oil price shocks are alike: disentangling demand and supply

shocks in the crude oil market.” American Economic Review, 99(3), 1053-1069.

Kormilitsina, A. (2011). “Oil price shocks and the optimality of monetary policy.”

Review of Economic Dynamics, 14(1), 199-223.

Leduc, S., Sill, K. (2001). “A quantitative analysis of oil-price shocks, systematic

monetary policy, and economic downturns.” Journal of Monetary Economics, 51(4),

781–808.

Mei, D.Z.; Gong; L.T. (2011). “The determinants of exchange rate regime in the

emerging economies.” Economic Research Journal(Chinese), 46(11), 73-88.

Mishkin, F.S. (2008). “Does stabilizing inflation contribute to stabilizing economic

activity?” NBER Working Paper, No.13970.

Mishkin, F.S. (2007). “Headline versus core inflation in the conduct of monetary

policy.” Proceedings of the Business Cycles, International Transmission and

Macroeconomic Policies Conference, Montreal, Canada, October 20.

Mishkin, F.S. (2010). “Inflation targeting in emerging market countries.” NBER

Working Paper, No.7618.

Mohanty, M.S.; Klau, M. (2005). “Monetary policy rules in emerging market

economics: issues and evidence.” In Monetary Policy and Macroeconomic

Stabilization in Latin America; Langhammer, R.J.; Springer: Berlin, Germany.

Smets, F.; Wouters, R. (2003). “An estimated dynamic stochastic general equilibrium

model of the Euro area.” Journal of the European Economic Association, 1(5), 1123-

1175.

Smets, F.; Wouters, R. (2007). “Shocks and frictions in U.S. business cycles: A

Bayesian DSGE approach.” American Economic Review, 97(3), 586-606.

Unalmis, D.; Unalmis, I.; Unsal, D.F. (2009). “On the sources of oil price fluctuations.”

IMF Working Papers, No. 285.

Wang, Y.Q.; Zhu, Q.G. (2015). “Energy price shocks, monetary policy and China's

economic fluctuations.” Asian-Pacific Economic Literature, 29(1), 126–141.

Woodford, M. (2003). “Interest and prices: foundations of a theory of monetary policy.”

Princeton University Press: New Jersey, USA.

Xie, P.; Luo, X. (2002). “Taylor rule and its empirical test in China’s monetary policy.”

Economic Research Journal(Chinese), 3, 3-12.

Zhang, W.L. (2009). “China’s monetary policy: quantity versus price rules.” Journal of

Macroeconomics, 31(3), 473-484.

Zhao, L.; Zhang, X.; Wang, S.Y.; Xu, S.Y. (2016). “The effects of oil price shocks on

output and inflation in China.” Energy Economics, 53, 101–110.

Zhong, C.P.; Sun, H.M.; Xu, C.S. (2010). “Credit constraint, credit demand, and

behavior of farmer’s lending.” Journal of Financial Research (Chinese), 11, 189-206